Back to skill

Security audit

resume-optimizers

Security checks for vulnerabilities and agentic risk

Overview

This resume tool is purpose-aligned, but users should know it sends resume text to Claude and writes a local HTML report containing resume details.

Install only if you are comfortable sending the full text of a resume, including contact and employment details, to Claude for analysis. Consider redacting sensitive fields before use, review the generated HTML before sharing it, and prefer pinned dependency versions for reproducible installs.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
scripts/requirements.txt:1
Finding

Unpinned Third-Party Dependencies

Content
View full analysis

Vulnerability Details

File Location: scripts/requirements.txt:1-3
Vulnerability Type: Unpinned and unnecessarily included third-party dependencies
Risk Level: Medium

Vulnerable Code

text
pdfplumber
fpdf2
anthropic

Technical Analysis

All third-party dependencies are specified without exact version constraints or integrity hashes. Consequently, separate installations of the same project can retrieve different package versions containing code that was not reviewed during this audit. This creates exposure to compromised upstream releases, malicious dependency updates, and unexpected security regressions.

The fpdf2 package also appears unused by the audited script. Retaining an unnecessary package increases the supply-chain attack surface without supporting the documented functionality.

This finding does not establish that any currently published version of these packages is malicious. The vulnerability is the absence of reproducible, integrity-verified dependency resolution.

Attack Path

  1. An attacker compromises the publishing account or release process of one of the listed packages, or otherwise causes a malicious version to become available through the configured Python package index.
  2. A user or deployment process runs pip install -r scripts/requirements.txt.
  3. Because no exact versions or hashes are specified, the installer resolves and downloads the malicious or vulnerable release.
  4. Package installation hooks or subsequently imported package code execute with the privileges of the process running the installation or resume optimizer.
  5. The malicious dependency can access data and resources available to that process, potentially including resume contents, environment variables, and generated reports.

Impact Assessment

Successful exploitation would execute dependency-controlled code with the privileges of the user or service account installing or running the Skill. The accessible scope could include loc ...[truncated 298 chars]

Remediation
View remediation

Remediation Suggestions

  1. Remove fpdf2 if it is not required by any supported execution path.

  2. Pin every direct dependency to an exact, reviewed version, for example:

    text
    pdfplumber==REVIEWED_VERSION
    anthropic==REVIEWED_VERSION
    
  3. Generate a lock file that includes reviewed transitive dependencies.

  4. Record and verify package hashes, and install with pip's --require-hashes option.

  5. Obtain dependencies only from an explicitly configured trusted package index.

  6. Integrate dependency vulnerability and provenance scanning into the release process.

  7. Regularly update pinned versions through a controlled process that includes review, testing, and regeneration of integrity hashes.

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
Findings (15)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill describes behavior that handles resume files, sends extracted content to a model, and produces output artifacts, but it does not declare any explicit tool scope or permission boundaries. Missing scope declarations can allow broader-than-expected access to environment data or file-writing behavior, increasing the chance of unintended file modification or data exposure during execution.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The trigger phrases are broad natural-language requests like '修改简历' and '优化简历' without clear boundaries or disambiguation rules. This can cause the skill to activate on unintended conversations, potentially processing sensitive resume data or generating files when the user did not explicitly intend to invoke this workflow.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill states that extracted resume content will be sent to Claude for evaluation, but it does not clearly warn users that personal data from resumes may be transmitted to an external model service. Because resumes commonly contain names, phone numbers, email addresses, employment history, and other sensitive personal information, lack of disclosure creates a meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The title and entire evaluation guidance are written exclusively in Chinese, establishing a language-specific instruction set for the skill. There is no indication that the user can choose another language or that the Chinese-only constraint is required for a region-specific purpose, which can violate language/locale policy requirements.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill description does not disclose that full resume contents will be sent to an external LLM service for processing. Resumes commonly contain highly sensitive personal and employment data, so this omission prevents informed consent and can lead to unintended third-party disclosure.

Content

No source excerpt is available for this finding.

Ssd 4

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The prompt instructs the external model to perform an exhaustive review of the full resume and reproduce detailed information, increasing the amount of sensitive user data processed and reflected back. In the context of a resume-optimization skill, this makes privacy exposure worse because the skill handles inherently personal documents and sends them wholesale to a third party.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The prompt explicitly asks for original resume descriptions and detailed highlights to be included in the generated report, which can unnecessarily duplicate sensitive user-provided content. This increases exposure in both the external API processing stage and the resulting HTML artifact, which may be shared or stored insecurely.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The code sends the extracted resume text directly to an external API without any user warning or confirmation at the point of transmission. Because resumes often include names, contact details, work history, and other personal data, silent transfer to a third party creates a meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
97% confidence
Finding

The dependency pdfplumber is specified without a version pin, which makes builds non-reproducible and allows future installs to pull in unexpected or vulnerable releases. In a skill that processes resumes and likely handles user-supplied PDF files, dependency drift increases supply-chain and reliability risk even if no specific vulnerable version is identified here.

Content

Scanner excerpt · scripts/requirements.txt (reported line 1)May include surrounding context.

text
pdfplumber
fpdf2
anthropic

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
97% confidence
Finding

The fpdf2 package is unpinned, so installations may resolve to different versions over time, including releases with breaking changes or security issues. Because this skill generates HTML/PDF-style output artifacts, uncontrolled dependency updates can introduce supply-chain risk and reduce reproducibility.

Content

Scanner excerpt · scripts/requirements.txt (reported line 2)May include surrounding context.

text
pdfplumber
fpdf2
anthropic

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
99% confidence
Finding

The anthropic SDK is unpinned, which means the environment may install an arbitrary current version, including versions later found vulnerable. This is more concerning here because the static analysis also notes known advisories for anthropic, and the skill likely interacts with external model services and local files, increasing the chance that a vulnerable SDK could affect sensitive resume data or local execution context.

Content

Scanner excerpt · scripts/requirements.txt (reported line 3)May include surrounding context.

text
pdfplumber
fpdf2
anthropic

Unverifiable Dependency: anthropic has 4 known advisory(ies) (CVE-2026-34450 (Claude SDK for Python has Insecure Default File Permissions in Local Filesystem ); CVE-2026-34452 (Claude SDK for Python: Memory Tool Path Validation Race Condition Allows Sandbox); CVE-2026-34450 (The Claude SDK for Python provides access to the Claude API from Python applicat) +1 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
93% confidence
Finding

The manifest includes anthropic without a pinned version even though the package has known advisories, so it is impossible to determine whether the installed release is vulnerable. In this skill context, an SDK used for model access may have access to prompts, user resume content, API credentials, and possibly local filesystem interactions, so an affected version could expose sensitive data or weaken sandbox/file protections.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

该技能的声明用途是简历评估与生成 HTML 报告,但代码还会主动从环境变量读取外部服务凭据。虽然这是调用第三方 API 的常见实现方式,但从“简历优化”这一业务目的本身看,凭据访问属于额外能力,且 manifest 未披露。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The system prompt explicitly instructs the model to respond in Chinese, and the script’s interface and generated report are also Chinese-specific. For a general resume evaluation tool, this is a natural-language locale constraint without any user opt-in or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

In the backend branch prompt, the model is again required to return results in Chinese only. This creates the same locale-policy issue because users are not given a language option and the restriction is not justified in the file.

Content

No source excerpt is available for this finding.